Voice synthesis device

Abstract

A voice synthesis device which includes a database configured to store a voice and a text corresponding to the voice and a processor configured to extract characteristic information and a tone of a first-language voice stored in the database, classify an utterance style of an utterer on basis of the extracted characteristic information, generate utterer analysis information including the utterance style and the tone, translate a text corresponding to the first-language voice into a second language, and synthesize the text, translated into the second language, in a second-language voice by using the utterer analysis information.

What is claimed is: 
     
       1. A voice synthesis device comprising:
 a memory configured to store speech audio and corresponding text in a database; and 
 a processor configured to: 
 extract characteristic information and a tone of a voice of a first-language speech audio stored in the database, wherein the characteristic information comprises a plurality of style items including a sex of the utterer, a pitch of the voice of the first-language speech audio, an utterance speed of the utterer, and an emotion of the utterer; 
 assign a weight value to each of the plurality of style items included in the characteristic information; 
 classify an utterance style of an utterer based on combinations of a style item and the assigned weight value corresponding to the style item; 
 generate utterer analysis information including the utterance style and the tone; 
 translate a text corresponding to the first-language speech audio to a second language; and 
 generate a second-language speech audio based on the translated text using the utterer analysis information. 
 
     
     
       2. The voice synthesis device of  claim 1 , further comprising a preprocessor configured to extract a waveform of the voice of the first-language speech audio, a frequency band of the voice of the first-language speech audio, and a power spectrum of the voice of the first-language speech audio. 
     
     
       3. The voice synthesis device of  claim 2 , wherein the processor is further configured to extract the tone and the characteristic information by using one or more of the waveform of the voice of the first-language speech audio, the frequency band of the voice of the first-language speech audio, and the power spectrum of the voice of the first-language speech audio extracted by the preprocessor. 
     
     
       4. The voice synthesis device of  claim 1 , wherein the processor is further configured to:
 extract characteristic information of a voice from each of a first plurality of sentences of the first-language speech audio; and 
 classify a corresponding utterance style for each of the first plurality of sentences based on the extracted characteristic information. 
 
     
     
       5. The voice synthesis device of  claim 4 , wherein the generated second-language speech audio comprises a second plurality of sentences which are generated by applying utterer analysis information corresponding to the first plurality of sentences respectively to the second plurality of sentences. 
     
     
       6. The voice synthesis device of  claim 1 , wherein the tone of the voice of the first-language speech audio is reflected in the generated second-language speech audio by applying the generated utterer analysis information. 
     
     
       7. An operating method of a voice synthesis device, the operating method comprising:
 extracting characteristic information and a tone of a voice of a first-language speech audio stored in a database, wherein the characteristic information comprises a plurality of style items including a sex of the utterer, a pitch of the voice of the first-language speech audio, an utterance speed of the utterer, and an emotion of the utterer; 
 assigning a weight value to each of the plurality of style items included in the characteristic information; 
 classifying an utterance style of an utterer based on combinations of a style item and the assigned weight value corresponding to the style item; 
 generating utterer analysis information including the utterance style and the tone; 
 translating a text corresponding to the first-language speech audio to a second language; and 
 generate a second-language speech audio based on the translated text using the utterer analysis information. 
 
     
     
       8. The operating method of  claim 7 , further comprising:
 extracting characteristic information of a voice from each of a first plurality of sentences of the first-language speech audio, and 
 classifying a corresponding utterance style for each of the first plurality of sentences based on the extracted characteristic information. 
 
     
     
       9. The operating method of  claim 8 , wherein the generated second-language speech audio comprises a second plurality of sentences which are generated by applying utterer analysis information corresponding to the first plurality of sentences respectively to the second plurality of sentences. 
     
     
       10. The operating method of  claim 7 , further comprising extracting a waveform of the voice of the first-language speech audio, a frequency band of the voice of the first-language speech audio, and a power spectrum of the voice of the first-language speech audio. 
     
     
       11. The operating method of  claim 10 , wherein:
 extracting the tone and the characteristic information is performed by using one or more of the waveform of the voice of the first-language speech audio, the frequency band of the voice of the first-language speech audio, and the power spectrum of the voice of the first-language speech audio. 
 
     
     
       12. The operating method of  claim 11 , wherein the tone of the voice of the first-language speech audio is reflected in the generated second-language speech audio by applying the generated utterer analysis information.